Tuesday, October 8, 2013

All the Presidents' Men

A few years back, Jazzbumpa and I disagreed endlessly about whether some particular graph showed an exponential pattern or not, or more like, or less like. In the middle of all that, one day I came across a post at Historinhas. Marcus Nunes had taken a graph and divided it up into sections. He showed that one of those sections fit an exponential pattern, and the others didn't. I learned something that day. And I've been going to Marcus's ever since.

Oh, and Jazz was probably right.

At Historinhas, in a post called The “Great Moderation”=the “Great Stagnation”? Not according to ‘phase diagrams’, Marcus Nunes evaluated Jazzbumpa's A New Look at Real GDP and A New Look at Real GDP - Part 2.

Marcus turns thumbs down on one of Jazz's graphs, in part because Jazz evaluates the data in 8-year grabs -- an 8-year moving average, and like that. Marcus says, "I don´t really understand his choice of periods. Why 8 years?"

But Jazz refers to the data trend over presidential terms. "The last data point in the term represents the performance of a given 8-year administration," he writes, "and the trend over the term can offer contrast to other administrations." I thought that was pretty clear.

Anyway, Marcus offers an alternative way to break the data into chronological packets. He uses common economic subdivisions -- the “Golden Age”, the “Great Inflation”, “Volcker Transition”, the “Great Moderation”, and the “Great Recession”. And he associates each time period with a different Fed Chairman.

I don't care for political subdivisions, myself. But that doesn't mean the economic subdivisions are better. For example, apart from the Volcker Transition, all of Marcus's categories describe not policies but results. The economy was golden, or inflationary, or moderated, or recessing: All of these are results.

Another objection: By definition, the "Great Moderation" ended when the "Great Recession" started, simply because the Great Recession was not moderate. That doesn't mean the policies differed in the two periods. I guess Marcus argues that the policies differed. I'm not yet convinced.

If Fed Chairmen are more useful to economic analysis than Presidents, then I ought to be able to find evidence of policy differences that correspond to changing economic conditions. That's where I'm going in this post.

The dates of Marcus's subdivisions vary some in his post (as perhaps they do in mine). I'll go with the dates on his "genie" graphs:

1955-1969MartinGolden Age
1970-1977BurnsGreat Inflation
1987-2005GreenspanGreat Moderation
2006-2013BernankeGreat Recession

I'll use annual data and I'll use Marcus's dates unchanged, except I'll stop at 2012 instead of 2013 (because we don't have the data yet for all of 2013).

What shall we look at? How about base money, for starters. That's a variable that is controlled by the Fed. As opposed to NGDP and RGDP and inflation and moderation, which are results of policy. Results of policy and other factors.

I got annual AMBSL (St. Louis Adjusted Monetary Base) numbers from FRED and made a graph showing "percent change from previous year" with each subdivision shown in a different color.

Graph #1: Annual Percent Change in AMBSL, 1955-2012

But the base money increase of the Bernanke years dwarfs all else. So if you want to see what Bernanke has done, the graph is useful. But if you want to see anything else, it's not. So I eliminated the data for the years after 2007, leaving poor Ben Bernanke with only two years on my graph:

Graph #2: Annual Percent Change in AMBSL, 1955-2007
Now at least you can see the "action" in base money since 1955. You can see, for example, that base-money growth was nearly as fast in the latter years of William McChesney Martin's chairmanship (blue) as it was under Arthur Burns (red).

You can see -- surprisingly -- that base-money growth under Paul Volcker (orange) was higher even than under Arthur Burns. And that under Greenspan (green), base money growth was not much lower -- though the up-and-down variation under Greenspan
was near twice what it was under either Volcker or Burns. So much for moderation!

I went back to FRED and recreated my graph there, as a visual check on what I was doing in Excel:

Graph #3: Annual Percent Change in AMBSL, 1955-2007

Finally, I did a new graph in Excel showing the whole 1955-2007 1948-2007 period, and put a Hodrick-Prescott trend on it -- the red line that looks like an old Volkswagen:

Graph #4: Annual Percent Change in AMBSL (1955-2007) and Hodrick-Prescott (Lambda=100)

Here's one more look at the trends in AMBSL growth, this time from my son Jerry's "Curve Fit" program:

Graph #5: Jerry's Computer-Fitted Trend Lines

So now, the main question: How and when did policy change? More specifically: Did policies change with the transition from Greenspan to Bernanke? That's the key question, because we had the financial crisis.

According to Marcus, things were great under Greenspan, and Bernanke's policies were wrong, wrong, wrong. But according to my last two graphs (with trend lines) the growth of base money began slowing in the late 1980s or early 1990s. If that's the case, maybe Bernanke did nothing worse than stick to Greenspan's plan, and get left holding the bag.

// The Excel Spreadsheet to view or download from Google Drive.


João Marcus said...

Art, what you showed is that base money is NOT an indicator of the stance of monetary policy!


The Arthurian said...


Thanks, Marcus.

Jazzbumpa said...

Art -

Busy week. I missed this post until this morning. I wondered why you were focusing on base money, and it looks like Marcus has a point.

So - is base money an indicator of anything? And if so, what? And if not, why not?

Always more questions.


The Arthurian said...

Hi Jazz.

Base money, because that's the thing the Fed creates and issues at will, to influence the economy.

(People say the Fed controls interest rates; but that control depends on interaction with the private sector. Anyway it is base money that they issue, in order to "control" interest rates. Plus, as you wrote some time back, it looks like the Fed responds to interest rates, as opposed to setting them. In any event, they do things with base money to affect the economy.)

I thought the big arc shape of the trend is pretty interesting. Perhaps related to SRW's take on the baby boom.

Clonal has shown that the changes to base money (quantitative easing aside) are extremely small in comparison to the changes to money supply caused by Federal deficits. You'd think Milton Friedman would have pointed that out, but no.